When developers face graph problems like hierarchies, route finding, or social networks, they often reach for Neo4j or NetworkX. But for operational data with thousands of nodes—supply chains, org charts, navigation paths—a relational database can handle most tasks. The key is Recursive Common Table Expressions (CTEs), a SQL standard feature from 1999 rarely used today.

CTEs are temporary result sets within SQL statements, making complex queries readable by breaking logic into named sections. Recursive CTEs extend this by referencing themselves to iteratively generate rows until no new data emerges, effectively performing breadth-first search.

This article demonstrates graph traversal, pathfinding, and cycle detection using only standard SQL. Examples work with Postgres, Oracle, MySQL, or SQLite, though syntax varies slightly. For massive billion-node graphs, specialized tools remain necessary, but most business graph problems fit comfortably in existing relational databases using Recursive CTEs.